The output is the position of the finger tips, as well as the center of the palm. The input to the program was an image showing my hand. One of the requirements of the project was that the recognition be based solely on an object’s contour, without the use of active or passive markers.Īs a fun experiment, I decided to implement a hand recognition and tracking algorithm. The goal was to allow for recognition and tracking of custom objects as they were placed and moved around on the table’s surface. It makes use of a custom-built multi-touch table (40“). The following is the result of a relatively short (five week) research & development project for a company in Paris.
VOODOOPAD REQUIREMENTS SOFTWARE
I built the hardware and wrote the software (the hardware input detection and output generation, and the front-facing app). The project was submitted to a CompVis competition and was awarded the first prize (30 projects total). This can be used in medical meetings, or for doctors to explain and demonstrate things to patients. It allows medical data to be displayed on custom multi-touch installations. “An Exploration and Planning Tool for Neurosurgical Interventions”, IEEE Visualization Contest 2010, Honorable MentionĪnother project I created has a medical component, too, but is more user interaction centric.“Advanced Line Visualization For HARDI”, Bildverarbeitung für die Medizin (BVM) (2012).I have co-authored two papers on this and related visualization techniques: The image on the right shows an experiment of the same concepts, but with an added ambient occlusion effect. A temperature-inspired color coding is used to convey distance (red is closer than blue). The following image shows the original on the left, and my line art approach in the middle. You then select a region of the brain where you want to see the nerve fibers, e.g. when planning a surgical removal of a tumor, in which you don’t want to damage brain nerves that are practically invisible once you ‘cut a patient open’.
The input to the program is a diffusion MRI scan of a person’s brain. One such attempt is shown below, where I wrote an OpenGL-based prototype with the aim of helping neurosurgeons understand nerves in the human brain better. The main challenge in this field is to make vast sets of data visually comprehensible. I’ve done some work in the field of medical visualization.
VOODOOPAD REQUIREMENTS CODE
Originally developed by an external company, the code base grew to a massive size over time and needed some loving care.įor a short time in 2012, I helped fixing some of the lower-level issues and made the app more usable on the original iPad. develops the initially crowd-funded iPad application Taposé. The application also allows uploading app builds as they are created, and facilitates using crash report data. I worked on HockeyApp for Mac, HockeyApp’s companion app that helps developers integrate HockeyApp into their existing applications.
Distribute beta versions of their apps.Collect the best crash reports for iOS/Mac/Android.
VOODOOPAD REQUIREMENTS WINDOWS
HockeyApp is a platform that helps iOS, Mac, Android, and Windows Phone developers to: This included detection of the affine transformations between matches, and detection of optical differences (e.g. typo corrections). I also developed a prototype of an internal tool that employed computer vision techniques to map bitmaps generated from their database (vector graphics) of comics to the original comic and page.
0 kommentar(er)
